Foundation Publishes 100,000 Patient Study on "Total Mutational Burden" (TMB)

Very early on, before sequencing 200 or more genes in tumors was really commercialized, it occurred to me that from a business strategy perspective you would need to come up with a clinical reason for assessing 200+ genes.   Even today there are a very finite number of effective drugs for targeted mutations (EGFR, ALK, BRAF, etc), not anywhere near 200-500.   Adding referral to clinical trials will have limited impact; although tallies count "hundreds, or even thousands" of oncology clinical trials, they are likely counting things like phase I drug dosing arms.

Two approaches to making use of all 500 genes are being discussed.

• One use would be to make very complex in silico models of tumor cell biology.  Companies like Cellworks and Darwin Health are exploring this.  I think of this as a growing industry for "digital genomics."  
• Another use is finding something about those 500 sequenced genes that can be assessed and applied clinically -- but requiring the exponential jump in gene data and unrelated to the small number of approved drug-gene pairings.  
• An approach here is Total Mutational Burden (TMB), and Foundation Medicine has just published a large new study.

Note that while either or both of these areas may prove to be clinically effective and important, you would come across these kinds of ideas, just by working backwards, Sherlock Holmes-style, from assuming you already have built the 500- or 1000-gene test and want to reverse-engineer applications for it.  As an if-then statement;  if one of these ideas proves important, then the market for 500 gene tests would rise while not depending on the trickle of a few new individual FDA-approved gene-drug pairs.

Total mutational burden is being cited as a potential biomarker for the effectiveness of immuno-oncology checkpoint drugs.   Chalmers et al. report the results from 100,000 patient studies conducted in a wide range of tumors from the FMI archives (Genome Medicine, Open Access, here.) FMI finds that:

• They confirmed the previous understanding that lung cancer and melanoma are most likely to have rampant mutations, but found that rampant mutations also occurred at lower rates in many cancers.
• You get most of the information for 0.5MB DNA; an exome confirms the high or low mutation burden you already low, whereas accurate assessment falls off below 0.5MB.
• PMS2 mutations tend to be associated with rampant mutations, as is aging.

One proposed use of this type of data is predicting immunotherapy outcomes with a biomarker (TMB) orthogonal to markers like PDL1.   Last December, MolDX providing coverage of comprehensive genomic profiling in several additional cancer types (colon, melanoma, ovarian) based in part on the contributory value of TMB to therapy choices.

The authors also used Cancer Genome Atlas (TCGA) data.  Coverage at Genomeweb, here.

Update May 9:  FMI emphasized the growing role and value of TMB in its May 9 investor call, here.
Update October 6:   For a UCSD paper on TMB clinical impact (Khagi et al.) here.
______

For AACR news with a re-analysis of Opdivo data using TMB, BiopharmaDive, here.

For an earlier FMI paper on 63,220 tumors, see Hartmaier et al., Genome Med 9:16, February 2017, here.